Chips and Mobile Stocks Taking It on the Chin

Ars Techica reports: In recent days, mobile handset makers have been issuing warnings that
the cellphone market won't match previous growth projections. Samsung told Reuters
that "the market could post a single-digit or even negative growth"
next year, due to a massive slowdown in consumer demand. Samsung's
predictions come on the heels of dire warnings by Nokia that worldwide
sales of handsets and telecommunications equipment should actually
decline from 2008 to 2009.

Nokia's warning
on the 14th kicked off a dive in the stock prices of Asian handset
manufacturers, and the overall grim outlook for the wireless industry
is almost certainly a major factor behind the absolute beating that
semiconductor stocks have taken in recent weeks. The embedded
chipmakers who make the semiconductors that power mobile devices have
been hit the hardest, with Linear Technology, Marvell, Qualcomm, and
many others all down in the mid-double-digits off of their most recent
summer peaks.

Contra Dave Manners in a recent blog post,
you really don't have to resort to theories about the lack of
fundamental R&D in the semiconductor business to account for
price-to-earnings (P/E) ratios that are now in the teens(!). The low
P/Es simply indicate that investors have drastically cut their
expectations of future earnings growth, and for good reason. Even the
normally bullish Semiconductor Industry Association also predicted a contraction in demand next year as one market after another implodes.

Even with US consumers in full retreat, China was supposed to be a big
growth market for handsets and other "post-PC" computing devices, but
now fears that the country won't be able to sustain its voracious
double-digit growth rate (despite any announced stimulus plans) have
torpedoed expectations about Chinese demand. Demand growth for
semiconductors of all types in all markets will be negative next year,
and that reality will hurt chipmakers' margins more than most.

The pipeline

It takes a relatively long time to design chips, develop fab
technology, and actually build fabs, and the end result is that the
semiconductor industry as a whole has a fairly deep manufacturing
pipeline that requires it to estimate demand years out.

As we've seen in the flash memory business,
if the industry misses its projections and builds too much fab
capacity, the resulting supply glut is murder on profit margins. If the
chip industry had global demand growth through 2009 baked into its
estimates years ago as it began allocating resources for the current
45nm node (where "resources" includes everything from R&D up
through actual fab construction), then even if the demand curve levels
off, it means trouble. In 2009, that curve may actually trend downward,
hence P/Es at record lows.

In other words, very low P/Es are a pretty rational expectation at
this point if the market thinks that the pipeline for the semi industry
as a whole has "mispredicted a branch," so to speak. If global demand
fails to materialize next year, it means huge losses in either margins
or wasted capex (if fabs are shut); those predicted losses are are
being priced in today. 

Also on Ars Technica:

Blockbuster's New Set-Top Box

Cyberchondriac Searches

Acidic Oceans

---

Laura Rich is a co-founder of Recessionwire, which provides news, advice, perspective and humor about the recession and the recovery.